[email protected] -...
TRANSCRIPT
Conflict studies on urban roads International Course on Transportation Planning and Road Safety,
IIT Delhi, India, December, 2016
Christer Hydén
Professor emeritus in Traffic Engineering
2016-12-16 IIT Delhi 1
Conflicts as a surrogate to accidents
•There is a strong need for a systematic evaluation/assessment
of ”everything that is done or planned to be done”
•Without this we end up with safety problems that we should
have detected long ago. Compare the Swedish “zebra crossing
case”….
•…. And sub optimise our common resources
IIT Delhi 22016-12-16
The Swedish “zebra crossing case” (2)
Driver behaviour at zebra crossings, before making
yielding for pedestians mandatory 1st of May 2001
• Few drivers yield for pedestrians
• Injury accidents have increased
Conclusion
• There is a need to force drivers to yield for
pedestrians
• A new law is the solution!?
• Introduced 1st of May 2001
IIT Delhi 32016-12-16
What about other measures?
• What about retro-reflective poles? • (Effect on injury accidents: -7% (-22 -- +12))
• And what about all IT-solutions, like Electronic Stability Control…
• (Effect on ESC-related injury accidents: Around minus 50%)• (Effect on pedestrian and bicycle accidents: +4% (-12;+22)
• …and the new Driver Fatigue early warningsystem….
• ????
• ….and ISA (Intelligent Speed Adaptation)• (Rather popular, ineffective)• (Effect on
• ….and Speed Limiter in cars• (Not very popular. Effective)
IIT Delhi 42016-12-16
What happens???
A lady on her bike gets almost killed – peacefully on her way to go shopping…….
(VIDEO)
IIT Delhi 52016-12-16
What happens here???
What actually happened?
Why did it happen?
What about speed adaptation?
Any behavioural mistakes?
Any perceptional mistakes?
How can we avoid a similar situation?
How can we test?
IIT Delhi 62016-12-16
Why do we need surrogates to
accidents?
• Accidents are rare events and are therefore associated with random variation
• Not all accidents are reported and the level of reporting is unevenly distributed
• The behavioural or situational aspects of the events are not covered by police accident data
• Accident analysis is a desk tool, not a field tool
• We need links between accidents and behaviours
IIT Delhi 72016-12-16
Lots of convincing arguments. Still…
• Standardised use of conflicts or near accidents is veryrare
• It is only recently that the topic became interesting
• And much thanks to the IT-(r)evolution
• The big thing in Automotive research is Naturalistic Driving.
• Millions of data from car driving – but what should it be used for??
• And how??
IIT Delhi 82016-12-16
Serious Conflicts
Slight Conflicts
Potential Conflicts
Undisturbed
passages
Accidents
Hydén, 1987
We need a link between accident
and behaviour/interaction
2016-12-16 IIT Delhi 9
Relation between different kinds
of elementary events
• Very few very serious events
• Very many events with low severity
2016-12-16 IIT Delhi 10
Traffic Conflict Techniques in general
•In theory there are very many ways of defining
conflicts
•In practise there are very few
•In real life there are almost none
•Reliability and validity studies are missing
•One (partly) exception is the Swedish Traffic Conflcits
technique
2016-12-16 IIT Delhi 11
Definition of a conflict
A situation where two road users
approach each other in time and
space to such an extent that a
collision is imminent if their
movements remain unchanged.
• There are always only two road
users primarily involved
2016-12-16 IIT Delhi 12
This definition is universal
It is valid in all situations
where two road users
meet in a way that it
satisfies the basic criterion
of collision course
2016-12-16 IIT Delhi 13
ICTCT
Past:
• International Committee on Traffic Conflict Techniques
Now:
• International Cooperation on Theories andConcepts in Traffic Safety
http://www.ictct.org
2016-12-16 IIT Delhi 14
The Swedish Technique
2016-12-16 IIT Delhi 15
Two basic concepts
–Time to Accident (TA)
–Conflicting Speed (CS)
2016-12-16 IIT Delhi 16
TA = Time to Accident
The time that remains from one of the road
user have started an evasive action, until a
collision would have ocurred if the road
users had continued with unchanged speeds
and directions.
TA is calculated based on D and CS.
D= Distance to the collision point
CS= Conflicting Speed
2016-12-16 IIT Delhi 17
Conflicting speed (CS)
The speed (km/h or m/s) of the relevant
road user just prior to the evasive
maneouvre
2016-12-16 IIT Delhi 18
2/ Comparison of conflicts and
accidents regarding their TA and CS
values
2016-12-16 IIT Delhi 19
Conflict diagram
0
10
20
30
40
50
60
70
80
90
100
110
120
130
0 1 2 3 4 5 6
TA-value(sec)
km
/h
0
10
20
30
40
50
60
70
80
mp
h
Conflicting speed, CS (kph)
2016-12-16 IIT Delhi 20
Definition of alternative severities and
severity zones for testing-
5 zone alternatives
From: Hydén 19872016-12-16 IIT Delhi 21
Time to Accident (TA)
Conflicting Speed (CS)
Definition of alternative severities and
severity zones for testing –
Main criteria
1/ The risk in terms of accident to conflict
ratio should increase continuously from the
”lowest” class.
2/ Accident severity, i.e. number of fatal and
serious injuries in relation to the total
number of injuries should increase with
class.
2016-12-16 IIT Delhi 22
Car – BicycleTA and CS
accidentsconflictsFrom: Hydén 1987
2016-12-16 IIT Delhi 23
Car – PedestrianTA and CS
conflicts accidentsFrom: Hydén 1987
2016-12-16 IIT Delhi 24
ALT.DEF 2 fulfilled the criteria best
For definition of a serious conflicts
0,5 seconds margin was added
0
10
20
30
40
50
60
70
80
90
100
110
120
130
0 1 2 3 4 5 6
TA-value(sec)
kp
h
0
10
20
30
40
50
60
70
80
mp
h
Non-serious
conflict
Seriousconflict
Conflicting Speed (kph)
2016-12-16 IIT Delhi 25
Severity of conflicts
…. is defined by its position in the TA-
CS- graph
2016-12-16 IIT Delhi 26
A serious conflict
…. is a conflict that ends up to the left of the
border line in the “TA-CS”-graph
2016-12-16 IIT Delhi 27
Definition of a Serious Conflict in
practise
The TA-Value can be calculated based on
estimates of distance and conflicting speed
d= Distance to potential point of collision
2016-12-16 IIT Delhi 28
VB= u.k. km/hd1=u.k.m
Car A Car B
VA= 36km/h d1=15m
2016-12-16 IIT Delhi 29
Car A brakes; would have collided with Car B
if no braking
Conflict diagram
0
10
20
30
40
50
60
70
80
90
100
110
120
130
0 1 2 3 4 5 6
TA-value(sec)
km
/h
0
10
20
30
40
50
60
70
80
mp
h
Conflicting speed, CS (kph)
x
1,0 sec
36 km/h
2016-12-16 IIT Delhi 30
The character of a serious conflict
• Characterised by suddeness and
harshness in action of at least one of the
involved
• Even possible to build training on those
aspects
• Road users say that ”they would never
like to be involved in such a situation”
2016-12-16 IIT Delhi 31
Primary road user
Those two directly involved in the
conflict
(always only two; no matter how
many who are colliding)
Secondary road user
A road user who indirectly have
had importance for causing the
conflict2016-12-16 IIT Delhi 32
Relevant road user..
The road user that is relevant
for the defining of TA and CS
values
2016-12-16 IIT Delhi 33
27 km/h55 km/h
4 m15 m
Example 2
2016-12-16 IIT Delhi 34
Relevant road user is
”decided” either by...
...the road user who takes evasive action if
only one does
...if both take evasive action then the one of
the two primarily involved road users
whose combination of TA and CS produce
the value with lowest risk
2016-12-16 IIT Delhi 35
Example 2
0
10
20
30
40
50
60
70
80
90
100
110
120
130
0 1 2 3 4 5 6
TA-value(sec)
kp
h
0
10
20
30
40
50
60
70
80
mp
h
Non-serious
conflict
Serious
conflict
Conflict Speed (kph)
X
(relevant road user)
55 km/h1,0 sek
27
km/h 0,5 sek
(x)
2016-12-16 IIT Delhi 36
First try…..• You will see some conflicts on video
• You shall estimate the distance to the potential collision
point from the moment one of the road users take
evasive action (braking, swerving, acceleration)
• You shall estimate the speed of the road user who take
evasive action
• You use the Speed-Distance graph to find out the Time
to Accident
• You use the TA-Speed graph to decide whether a
conflict is serious or not
• Fill in your data on a blank piece of paper and compare
with the ”true value”
2016-12-16 IIT Delhi 37
Personal scorings. Name:
Personal scorings True values
Conflict
no.
Confl
speed
Distance
to Coll. P. TA
Conflict
no.
Confl
speed
Distance
to Coll. P. TA
382016-12-16 IIT Delhi
Speed-Distance-TA-value
2016-12-16 IIT Delhi 39
Video conflicts - facit
Conflict Speed
(km/h)
Distance
(m)
Speed
(m/s)
Time to
Accident (s)
(m/m//s
1 15 4 4.2 1.0
2 35 12 9.7 1.2
3 20 2 5.6 0.4
4 55 12 15.3 0.8
5 20 3 5.6 0.5
6 15 2 4.2 0.5
7 20 3 5.6 0.5
8 45 25 12.5 2.0
Conflicts on the conflict video. From 5:08
2016-12-16 IIT Delhi 40
Conflict no. 1 – 15km/h, 4 meter, TA=1.0 sec
0
10
20
30
40
50
60
70
80
90
100
110
120
130
0 1 2 3 4 5 6
TA-value(sec)
kp
h
0
10
20
30
40
50
60
70
80
mp
h
Non-serious
conflict
Serious
conflict
Conflicting Speed (kph)
IIT Delhi 412016-12-16
x
Conflict no. 2 – 35km/h, 12 meter, 1.2 sec
0
10
20
30
40
50
60
70
80
90
100
110
120
130
0 1 2 3 4 5 6
TA-value(sec)
kp
h
0
10
20
30
40
50
60
70
80
mp
h
Non-serious
conflict
Serious
conflict
Conflicting Speed (kph)
IIT Delhi 422016-12-16
x
Conflict no. 3 – 20km/h, 2 meter, TA=0.4 sec
0
10
20
30
40
50
60
70
80
90
100
110
120
130
0 1 2 3 4 5 6
TA-value(sec)
kp
h
0
10
20
30
40
50
60
70
80
mp
h
Non-serious
conflict
Serious
conflict
Conflicting Speed (kph)
IIT Delhi 432016-12-16
x
Conflict no. 4 – 55km/h, 12 meter, TA=0.8 sec
0
10
20
30
40
50
60
70
80
90
100
110
120
130
0 1 2 3 4 5 6
TA-value(sec)
kp
h
0
10
20
30
40
50
60
70
80
mp
h
Non-serious
conflict
Serious
conflict
Conflicting Speed (kph)
IIT Delhi 442016-12-16
x
Conflict no. 5 – 20km/h, 3 meter, TA=0.5 sec
0
10
20
30
40
50
60
70
80
90
100
110
120
130
0 1 2 3 4 5 6
TA-value(sec)
kp
h
0
10
20
30
40
50
60
70
80
mp
h
Non-serious
conflict
Serious
conflict
Conflicting Speed (kph)
IIT Delhi 452016-12-16
(x)
Conflict no. 6 – 15km/h, 2 meter, TA=0.5 sec
0
10
20
30
40
50
60
70
80
90
100
110
120
130
0 1 2 3 4 5 6
TA-value(sec)
kp
h
0
10
20
30
40
50
60
70
80
mp
h
Non-serious
conflict
Serious
conflict
Conflicting Speed (kph)
IIT Delhi 462016-12-16
x
Conflict no. 7 – 20km/h, 3 meter, TA=0.5 sec
0
10
20
30
40
50
60
70
80
90
100
110
120
130
0 1 2 3 4 5 6
TA-value(sec)
kp
h
0
10
20
30
40
50
60
70
80
mp
h
Non-serious
conflict
Serious
conflict
Conflicting Speed (kph)
IIT Delhi 472016-12-16
x
Conflict no. 8 – 45km/h, 25 meter, TA=2.0 sec
0
10
20
30
40
50
60
70
80
90
100
110
120
130
0 1 2 3 4 5 6
TA-value(sec)
kp
h
0
10
20
30
40
50
60
70
80
mp
h
Non-serious
conflict
Serious
conflict
Conflicting Speed (kph)
IIT Delhi 482016-12-16
x
Old video conflicts for training - facitTime/nu.
(sec/#)
Speed
(km/h)
Dist
ance
(met
er)
TA
Sec.
Time (sec Speed
(km/h)
Distance
(meter)
TA
Sec.
4:24/18 25 6 0.9 (ALLV) 13:52/35 25 8 1.2 (ALLV)
4:55/19 15 2 0.5 (ALLV) 13:59/35 35 14 1.4 (bra)
5:29/31 25 5 0.7 (ALLV) 14:55/46 10 1.5 0.6
6:56/22 35 8 0.8 (ALLV)+ 15:30/49 45 8 0.6 (bäst)
9:13/9 25 8 1.8 --- 16:36/24 15 3 0.7
9:46/10 45 13 1.0 (ALLV) 18:55/39 30 10 1.2
11:00/28 40 12 1.1 (ALLV) 21:14/11 15 3 0.7 (bra)
11:22 /44 45 10 0.8 (ALLV) + 1:58 (ny del)
11:53
//48
5 0,5 0.4 (intress.) 3:30
(paus)
13:28/27 15 3 0.7 (ALLV) +
Konflikter på konfliktfilmen. Från 5:08
2016-12-16 IIT Delhi 49
Practical use of the conflict technique (1)
• Conflicts are recorded using trained observers.
• Basic training normally takes one week
• Each conflict is recorded on a separate recording data sheet
• It is only serious conflicts that are used for the prediction of the expected number of accidents. However, the recording should include some more conflicts (to be on the safe side)
• Video is normally used as a complement and for behavioural studies, counts, etc
2016-12-16 IIT Delhi 50
Practical use of the conflict technique (2)
• Behaviour that seems to be interesting based on the conflict studies can be followed up systematically by specially designed behavioural studies
2016-12-16 IIT Delhi 51
2016-12-16 IIT Delhi 52
Practical use of the conflict technique (3)
• The most important task for the
observer is to estimate TA, with the
help of CS and distance,
• .....and to fill in the sheet correctly
• .....and always describe the events
preceeding the conflict ”as complete
as possible”
• .....and never lose attention...
2016-12-16 IIT Delhi 53
Practical use of the conflict technique (4)
• Normally one observer; more if the intersection is big or it is heavy traffic
• Select times of the day/year when problems seem to appear (e.g. Problems in darkness...)
• Still; normally studies are made 3 times 2 hours a day, morning, lunch and afternoon
• Normally 3 to 5 days of observations
• ”Extreme weather” should be avoided
2016-12-16 IIT Delhi 54
The reliability and validity of
the Swedish Technique has
been carefully studied
2016-12-16 IIT Delhi 55
Reliability and validity
1/ Comparison of different observers ability to
score correctly
2/ Comparison of the evasive maneouvre in
conflicts and accidents
3/ Comparison of conflicts and accidents
regarding there TA and CS values
2016-12-16 IIT Delhi 56
Observer reliability• The ”hit rate” is normally around
80%
• The average size of the difference in absolute values was in a comparative study 0,28 seconds
• In almost half of these conflicts the difference was less than ± 0,2 seconds
• In 82% the difference was less than ± 0,4 seconds
2016-12-16 IIT Delhi 57
1/ Comparison of different observers
ability to score correctly
From: Hydén 19872016-12-16 IIT Delhi 58
Product validity
• Three different samples of accidents and serious
conflicts from different intersections in two cities
produced quite similar “conflict to accident”
ratios
• The average annual expected number of police
reported injury accidents was often shown to be
better predicted by conflicts than by accidents at
single locations
2016-12-16 IIT Delhi 59
Conversion factors –
serious conflicts to accidents
Car-car Car-car Car-pedestrian Car-bicycle
p= 2.8 10-5
11.9 10-5
33.9 10-5
2016-12-16 IIT Delhi 60
Process validity
1/Comparison of the evasive maneouvre
in conflicts and accidents
2/ Comparison of conflicts and accidents
regarding there TA and CS values
2016-12-16 IIT Delhi 61
1/ Type of evasive action in conflicts
and accidentsMotor vehicles vs motor vehicles
braking braking swerving acceleration
+swerving
Conflicts 77% 16% 5% 3%
Accidents 73% 15% 10% 2%
2016-12-16 IIT Delhi 62
2/ TA and CS for conflicts and accidents
• Based on conflict studies
• Based on analysis of in-depth data
from accidents – from the same
locations
2016-12-16 IIT Delhi 63
Car – BicycleTA and CS
accidentsconflictsFrom: Hydén 1987
2016-12-16 IIT Delhi 64
Car – PedestrianTA and CS
conflicts accidentsFrom: Hydén 1987
2016-12-16 IIT Delhi 65
Conflicts and accidents belong to the same
process, just with different degree of
seriousness (most often)
• Patterns are very alike
• Accidents have a TA-value that is approx
0,5 seconds smaller and a CS that is
approx 10 km/h higher than serious
conflicts
• Accidents represent a logical continuation
of the serious conflicts on a severity scale
2016-12-16 IIT Delhi 66
Main conclusion
Conflicts and accidents
belong to the same process,
just with different degree of
seriousness
2016-12-16 IIT Delhi 67
Analysis of conflict studies
• There are between 3.000 och 40.000 conflicts on each police-reported injury accident (Swedish conditions)
• We use these conversion factors ()between accidents and serious conflicts
• In principle the analysis is made as if it were accidents, with the adding that conflicts are also described in terms of events
• The Data Base CDBase can be used to automatise the analysis
• It produces predictions automatically as well as other statistics
2016-12-16 IIT Delhi 68
Conversion factors ( ) –conflicts per day to police reported injury accidents per year
Car-Car Car-Car Carl-Pedestrian
Car-Cyclist
= 2.8 10-5
11.9 10-5
33.9 10-5
2016-12-16 IIT Delhi 69
Antallet forväntede personskadeulykker:
a/ Fra konfliktstudiet: Bil – Bil: 0,24 ul. per år. Bil – Myk trafikant: 0,35 ul.
per år
b/ Fra ulykkesstatikken:Bil – Bil: 0,10 ul. per år. Bil – Myk trafikant: 0,50 ul.
per år2016-12-16 IIT Delhi 70
Example on an assessment of a conflict study